Local variations in crime seasonality
Matt Ashby
Crime seasonality
- Crimes vary seasonally
- Seasonality related to activity patterns
- Many types peak in summer
- Seasonality might vary spatially
But …
- Little spatial disaggregation
- Modifiable temporal unit problem
- Seasonal analysis needs circular statistics
- Few inferential statistics
This project
- disaggregate data from multiple cities
- no temporal aggregation
- inferential tests for circular data
Some tentative conclusions
- Seasonality more consistent across crime types than within cities
- Peak is often neither ‘summer’ or ‘winter’
- Most neighbourhoods have similar seasonality, with (interesting) outliers